Typology of Occurings

JOHANNA SEIBT
FREE PROCESS THEORY: TOWARDS A TYPOLOGY OF

OCCURRINGS
ABSTRACT. The paper presents some essential heuristic and constructional elements of
Free Process Theory (FPT), a non-Whiteheadian, monocategoreal framework. I begin with
an analysis of our common sense concept of activities, which plays a crucial heuristic role
in the development of the notion of a free process. I argue that an activity is not a type
but a mode of occurrence, defined in terms of a network of inferences. The inferential
space characterizing our concept of an activity entails that anything which is conceived of
as occurring in the activity mode is a concrete, dynamic, non-particular individual. Such
individuals, which I call ‘free processes’, may be used for the interpretation of much more
than just common sense activities. I introduce the formal theory FPT, a mereology with a
non-transitive part-relation, which contains a typology of processes based on the following
five parameters relating to: (a) patterns of possible spatial and temporal recurrence (auto-
merity); (b) kinds of components (participant structure); (c) kinds of dynamic composition;
(d) kinds of dynamic flow (dynamic shape); and (e) dynamic context. I show how these
five evaluative dimensions for free processes can be used to define ontological correlates
for various common sense categories, and to draw distinctions between various forms
of agency (distributed, collective, reciprocal, entangled) and emergence (weak, strong, as
‘autonomous system’ (Bickhard/Christensen)).
The study of processes and process-based theory formation is a sur-

prisingly short chapter in the history of ontology. Despite Aristotle’s
sophisticated investigations into change and interactive development, des-
pite Whitehead’s bold attempt at a comprehensive process metaphysics, the
traditional research focus in ontology has been on ‘static’ entities such as
objects (substances), properties, relations, and facts. Those contemporary
ontological schemes which are deemed ‘revisionary’ demote the primacy
of ‘substances’ merely to turn still to other types of ‘static’ particulars such
as tropes or states of affairs.
As I have shown elsewhere, current analytical ontology still abides by
the (about twenty) presuppositions that characterize the “myth of sub-
stance”, research paradigm that has been so dominant in ontology, and I
have argued that the ‘substance paradigm’ hampers ontological explana-
tions of identity, qualitative sameness, and persistence.1 In the following I
sketch some elements of ‘Free Process Theory’, a new (non-Whiteheadian)
process-ontology which, to my knowledge, abandons more substance-
ontological principles than any other revisionary scheme hitherto pro-
Axiomathes 14: 23–55, 2004.

© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
24 JOHANNA SEIBT
posed, including Whitehead’s. Even though Free Process Theory operates

with a minimum of constructional basics, the scheme is sufficiently versat-
ile to support a rich typology of processes, including process-ontological
truth-makers for our talk about allegedly ’static’ entities such as things and
stuffs. Free processes are a novelty in ontological category theory, with
Broad’s/Sellars’ ‘subjectless’ or ‘pure’ processes and Zemach’s ‘types’ as
their closest systematic cognates.2 Free processes are modeled on (subject-
less) activities, i.e. on the denotations of sentences such as ‘it is snowing’,
‘it is getting warmer’, ’the light is shining’ etc. In section 1 I discuss
Vendler’s and Kenny’s notion of an activity and argue, based on a look at
recent aspectological research in linguistics, that the well-known distinc-
tion between activities, accomplishments (performances), achievements,
and states, is not a distinction in occurrences per se but a distinction in
modes of occurrence, i.e. in the way in which we take occurrences to
occur. The observation that different lexical contents can be conceptually
‘packaged’ in different ways is crucially important for ontological category
theory, since it can be taken to show that individuality is not tied to spatio-
temporal boundedness. In section 2 I introduce the category of a free
process, an entity that is dynamic, concrete, individual, and automerous
or repeatable (i.e. non-particular!). In section 3 I give an overview over the
main constructional elements of the mereological system used to formulate
Free Process Theory. Free processes stand in part-whole relationships, but
the part-relation on free processes is not transitive and thus the axioms of
classical extensional mereology do not apply. Due to its non-transitivity
the part-relation on processes can be diversified with respect to the degree
of mereological embedding (parts, parts of parts etc.). This has a variety
of definitional advantages, some of which become apparent in section 4
where I present a tentative list of classificatory parameters for a general
typology of processes.
1. WHAT IS AN ACTIVITY ?
The task of ontology is to explore various ways of structuring the referen-

tial domain of a language.3 More precisely, an ontological theory has the
form of the quadruple M, TM , f, L: it specifies an assignment f which
correlates the elements of a class L of L-sentences with “truth-makers”:
structures of the domain of interpretation M as described by a domain
theory TM . The assignment function f (which remains mostly implicit)
should be chosen in such a way that it can be used to explain why L-
speakers are justified in drawing categorical inferences, a certain type
of material inferences, which define the meaning of the ultimate genera
FREE PROCESS THEORY 25
terms of L (e.g., ‘thing’, ‘property’, ‘person’ etc.). This explanation takes

the following form: via subsumption under a certain category with certain
category features, the material inference is turned into a formal inference.4
Furthermore, given that the structural descriptions of ontology are to serve
explanatory purposes, ontological categories – just like theoretical con-
cepts in science – need a model or canonical illustration to serve their
explanatory function. In physics, a water current or an ideal spring serve as
cognitive models for electrical current or an harmonic oscillator, respect-
ively. Similarly, in ontology substances are modeled on things, and monads
and Whiteheadian occasions on minds. An ontological model must be
familiar to L-speakers or, as I say, ‘founded’ in their agentive experience.5
According to the postulate of foundedness any new candidate for an
ontological category must be shown to have a model, i.e. to be directly
associated with a concept which L-speakers are agentively familiar with.
Instead of first introducing the category of a free process and then its model
I will here turn matters around and begin with reflections on our concept
of an activity. As we shall see, if we pay proper attention to our common
sense understanding of activities – i.e. to the categorial inferences licensed
by those sentences that we commonly accept as being about actitivies –
we find that we are already well familiar with the concept of recurrent,
non-determinate, dynamic individuals to be introduced in the next section.
1.1. ‘Action types’: an unsuccessful start

Aristotle may have been the first to distinguish between occurrence types
in terms of inferential and linguistic criteria, sorting occurrences into en-
ergeiai and kineseis.6 In 1949 G. Ryle revisited the Aristotelian distinction
and distinguished ‘try it’ verbs or ‘verbs of activity or process’ such as
search, kick, or treat from ‘got it’ verbs or ‘achievement words’ such as
find, score, or heal.7 Inspired by Aristotle and Ryle, yet independently
of each other, Z. Vendler and A. Kenny developed more differentiated
classificatory schemes of ‘action verbs’ or ‘action types’.8 Vendler ’s clas-
sification is essentially based on four criteria, which I dub the dynamicity
condition (C1), the unboundedness condition (C2), the distributivity condi-
tion (C3), and the homeomerity condition (C4). For example, verb phrase
‘V’ is an ‘activity verb’ iff all of the following four conditions are met
(where ‘N’ stands for a noun-phrase in L):
(C1) Dynamicity: ‘N is V -ing’ is a well-formed L-sentence.
(C2) Unboundedness: ‘x finished V -ing’ is not a truly applicable L-
predicate.
26 JOHANNA SEIBT
(C3) Distributivity: For every temporal interval [t], if ‘A V -ed during [t]’
is true then ‘A V -ed during [t ]’ is true for every period [t ] that is
part of [t].
(C4) Homeomerity: Any temporal part of the denotation α of ‘V ’ is of the
‘same nature’ as the whole of α.
Vendler uses these four conditions to set up a fourfold division of ‘action
verbs’ and ‘associated time schemata’:
TABLE I
Activity verbs: run, walk, swim, push State verbs: have, possess, like, hate,
etc. desire, want, dominate, rule etc.
• dynamic • not dynamic
• unbounded • unbounded
• distributive • (strictly) distributive
• homeomerous • homeomerous
Time schema: ‘N was V -ing at t’ is true Time schema: ‘N V -ed between t1 and t2 ’
means that t is on a time stretch is true means that at any instant between
throughout which N was V -ing t1 and t2 N V -ed
Accomplishment verbs: paint a picture, Achievement verbs: start, reach the
build a house, grow up, recover from summit, win the race, be born/die, find,
illness, run-a-mile etc. recognize etc.
• dynamic • not dynamic, instantaneous
• bounded • unbounded
• not distributive • (trivially) distributive
• anhomeomerous • (trivially) homeomerous
Time-schema: ‘N was V -ing at t’ is true Time schema: ‘N V -ed between t1 and t2 ’
means that t is on the time stretch in is true means that the instant at which N
which N V -ed V -ed is between t1 and t2
Kenny champions a threefold classification into ‘states’, ‘activities’ and

‘performances’ (i.e., roughly speaking, a class comprising Vendler’s ac-
complishments and achievements). Like Vendler he operates with a mix-
ture of inferential and linguistic criteria, but pays greater attention to verb
forms than Vendler. For activities, for example, he adds an additional in-
ferential criterion familiar from Aristotle’s characterization of energeia,
which I call the completeness condition9 :
(C5) Completeness condition: ‘N is V -ing’ implies ‘N has V -ed’.10
Verbs which license the entailment specified in the completeness condition
Kenny calls ‘activity verbs’, those which fail to license the entailment
but fulfill the dynamicity condition (C1) are ‘performance verbs’. “For
instance, if a man is building a house, he has not yet built it; if John is
deciding whether to join the army, he has not yet decided to; if Mary is cut-
ting the cake, she has not yet cut it.”11 The idea behind Ryle’s, Vendler’s,
and Kenny’s approach is to distinguish types of verb denotations by distin-
guishing types of verbs. However, as has been pointed out in the literature,
there are a number of difficulties with the suggested classification.12 The
following two shortcomings are decisive. (a) If at all, the classifications
work for whole sentences rather than for ‘verbs’; not verbs (verb phrases)
but whole sentences (whole predications) carry the different inferential
roles that dovetail with ‘action types’.13 (b) While Vendler and Kenny
thought that lexical meanings of verb phrases could be sorted into four
(or three) types of occurrences, the non-lexical, aspectual meaning of verb
phrases plays an essential role for any such classification. Aspectual mean-
ings or ‘verbal aspects’ are expressed by special morphological elements
(‘verbal aspect markers’) or periphrastically. The difference in aspectual
meaning accounts, for instance, for the difference in meaning between the
sentences Tom crossed the street and Tom was crossing the street.14 Vendler
exploits aspectual meaning in formulating the ‘dynamicity condition’ (C1),
but fails to observe that changes in aspectual meaning systematically affect
his classification; for example, sentences with ‘accomplishments verbs’ in
the continuous form, such as Tom is crossing the street, support an ‘activity
reading’ rather than an ‘accomplishment reading.’ Kenny makes extensive
use of aspectual information (see below) but takes it to be strictly linked to
lexical information. Both classifications thus neglect the ‘phenomenon of
type shift’, i.e. the fact that a shift in the aspectual meaning of a sentence
can effect a shift in the occurrence type denoted by the sentence.15
1.2. Aktionsarten and verbal aspect: aspectological lessons

Vendler’s and Kenny’s project received little attention in contemporary on-
tology but was followed up in linguistic research on verb semantics, partly
within the field of formal semantics,16 partly constituting a new research
area that nowadays is labeled ‘aspectology.’ Aspectology is concerned
with information about the dynamic properties of the denotations of verb
phrases or sentences (predications) in general i.e. both with (i) aspectual
meaning or verbal aspects as well as with (ii) lexically coded information.
Occasionally aspectologists use the term ‘aspectual meaning’ to cover both
(i) and (ii), but I will use that term in the narrow sense of (i) and otherwise
speak of ‘processual information.’
There are two general aspectological research strategies to be
discerned.17 The older, ‘lexical strategy’, as one might call it, largely
pursues the Vendler/Kenny project of trying to sort verb phrases or pre-
dications into various occurrence types or ‘Aktionsarten’.18 Unlike Vendler
and Kenny, however, proponents of the lexical strategy are careful to define
28 JOHANNA SEIBT
‘Aktionsarten’ in terms of purely lexical criteria, i.e. in terms of possible

adverbial modifications or inferences engendered, and without reference
to aspectual information expressed by the English Continuous form or
Perfect tense. There are various systems of Aktionsarten, mostly set up
by means of feature combinations of a three binary parameters such as
‘+-dynamic’, ‘+-control’, ‘+-telicity.’ The resulting classifications of oc-
currence types differ somewhat and are differently diversified, but in each
case they include Vendler’s criteria (C2) through (C4) and, fairly consist-
ently, reproduce his four action types.19 Most significant for our purposes
is the fact that all systems of ‘Aktionsarten’ postulate a type of dynamic
and non-development or ‘atelic’ situations or ‘eventualities’, but have con-
spicuous difficulty in defining (in purely lexical terms!) the ‘dynamicity’ of
such situations. While there are ample criteria for the ‘atelicity’ and ‘teli-
city’ of a predication, i.e. for the difference between predications denoting
non-developmental or developmental dynamic situations, the distinction
between ‘static’ and ‘dynamic’ situations poses a genuine problem.20
The ‘lexical strategy’ not only falls short of making proper sense of
‘dynamicity’, it also has little to say about the phenomenon of ‘type
shift’ mentioned above. Both of these difficulties can be successfully ad-
dressed once ‘processual information’ about occurrence types is taken to
be conveyed by the ‘interaction’ or composition of a predication’s lexical
meaning and aspectual meaning. This is the second and currently favoured
strategy of aspectological research, where aspectual meaning plays an
integral part in determining the occurrence type denoted by a predication.21
Proponents of the ‘compositional strategy’ claim that the lexical mean-
ing of predications provides the schema of a dynamics, a temporal exten-
sion structure (“phase structures”, “temporal schema”), which is further
‘operated’ on by the processual information of verbal aspects. Of primary
importance in this regard is the contrast between the perfective aspect and
the imperfective or progressive aspect, respectively. The precise contribu-
tion in processual meaning provided by the perfective and imperfective
is apparently not easily stated. Linguistic characterizations of this differ-
ence exhibit considerable conceptual strain and frequently resort to spatial
metaphors: the perfective presents an occurrence ‘from the outside’, as a
‘bounded unit’, while the imperfective presents an occurrence ‘from the
inside’; the imperfective is said to present an occurrence in the way in
which a parade is experienced from the point of view of a “person march-
ing along in a parade” who is “in the middle of the parade and unable
to oversee . . . the full length of the parade”, while the perfective presents
it like the parade is experienced “from the point of view of a spectator
who, from an elevated vantage point, can oversee the entire extension of
the parade”.22 Elsewhere I have argued that the best sense we can make of
these descriptions is in terms of dynamicity: the perfective presents the fac-
tual content of an occurrence; the imperfective highlights the dynamicity of
the occurrence or presents the situation as occurrence; and the progressive
adds to the information that the situation occurs at the reference point of
the utterance.23
In order to formalize the difference between perfective and imperfective
‘focus’ of an occurrence, proponents of “selection theories” of proces-
sual meaning represent temporal extension structures (phase structures) by
means of two elements: the situation boundary and the transition (‘phase’)
between the boundaries. For example, a verb phrase may be lexicalized
with the temporal extension structure ‘[(τ )φ(τ )]’ (a phase and unspecified
boundaries) or with ‘[τ ]’ (just the boundary) or with ‘[φ]’ (just the phase)
or ‘[φτ ]’ (phase and specified boundary) etc. As recognizable from these
examples, four of these phase structures dovetail with Vendlerian action
types. The function of the imperfective and perfective aspect is to select for
denotation different elements of a phase structure: the imperfective selects
for the phase and the perfective selects for the boundaries.
However successful such compositional theories of processual meaning
might be as linguistic semantics, they are mainly theories of denotation
and tell us little about the conceptual role of processual information bey-
ond topological implications.24 But they furnish an important lead for the
clarification of our concept of an activity by treating aspectual information
as the decisive factor in the classification of the occurrence types denoted
by a predication. As shown in the following subsection, the differences in
processual information that led Vendler and Kenny to stipulate different
occurrence types can be articulated as differences in complex aspectual
information. The ‘viewpoint’ of the imperfective (progressive) aspect is
inferentially particularly fertile: all that matters about activities (and the
other three Vendlerian action types) can be expressed in terms of different
implicational roles of sentences with imperfective (progressive) aspect.
1.3. Activity as mode of occurrence

Unlike Vendler, Kenny included among the criteria for his ‘action types’
also inferential schemata involving English verb forms with aspectual dif-
ferences (cf. for example condition C5 above). To my knowledge, neither
the ontological nor the aspectological discussion has paid any attention
to Kenny’s inferential characterization of action types, mainly because
these inferences can be shown to fail for particular instances of the verb
types they are supposedly defining. But once we dispense with the idea
that occurrence types are fixedly denoted by verbs, Kenny’s observations
30 JOHANNA SEIBT
about aspectual implications can be put to good use. Instead of claiming

with Kenny that activity verbs are those that fulfill the schema ‘N is V -ing
implies N has V -ed’ (only to discover that for almost any verb we can
think up circumstances where such an implication does not hold), we will
take the implication as an indication that a certain occurrence denoted by
a predication is conceived of as an activity. Instead of claiming that run
or read are activity verbs in any context, and run a mile or read through
are accomplishment verbs in any context, etc., we need to concentrate
on the inferential role of a predication in a certain context and say that
a predication denotes a certain type of occurrence just in case it fulfills,
within that context, a certain inferential role.
More concretely, I suggest the following redefinition of Vendler’s four
action types in terms of sets of implications involving changes in aspect
or ‘aspectual implications’ for short; in formulating these implications I
take my bearings from Kenny, replacing English verb forms with gen-
eral, language-independent references to predications With progressive
aspect (‘prog(P )’), habitual aspect (‘hab(P )’), neutral aspect (‘neut(P )’),
and in the form of the experiential or resultative perfect (‘e-perf(P )’ and
‘r-pref(P )’).25
[M1] A sentence S with verbal aspect VA and predication P denotes a state iff
for VA = r-perf: r-perf(P ) ⇒ neut(P ) & ¬(r-perf(P ) ⇒ prog(P ))
and for VA = e-perf: ¬(e-perf(P ) ⇒ neut(P ))
and for VA = neut: ¬(neut(P ) ⇒ e-perf(P ))
[M2] A sentence S with verbal aspect VA and predication P denotes an accomplish-
ment/development iff
for VA = r-perf: r-perf(P ) ⇒ ¬prog(P ))
and for VA = e-perf: e-perf(P ) ⇒ ¬prog(P ))
and for VA = neut: neut(P ) ⇒ hab(P )
and for VA = prog: prog(P ) ⇒ ¬r-perf(P ) and ¬e-perf(P )
prog(P ) ⇒ r-perf + prog(P ) & e-perf + prog(P )
[M3] A sentence S with verbal aspect VA and predication P denotes an activity/process iff
(i) for VA = r-perf: r-perf(P ) ⇒ (prog(P )) ∨ ¬prog(P ))
(ii) and for VA = e-perf: e-perf(P ) ⇒ (prog(P )) ∨ ¬prog(P ))
(iii) and for VA = neut: neut (P ) ⇒ hab(P )
(iv) and for VA = prog: prog(P ) ⇒ r-perf(P ) & e-perf(P )
prog(P ) ⇒ r-perf + prog(P ) & e-perf + prog(P )
[M4] A sentence S with verbal aspect VA and predication P denotes an achievement/result
iff
for VA = r-perf: ⇒ (r-perf(P ) ⇒ neut(P ))
and for VA = e-perf: ¬(e-perf(P ) ⇒ neut(P ))
and for VA = neut: neut(P ) ⇒ r-perf(P ) & e-perf(P )
Since the clauses of these four definitions are conjuncted, an occurrence

type is defined via an entire network of aspectual implications. A sentence
like Tom is mending his shirt is thus ambiguous, for example. It denotes
an accomplishment just in case we take it to imply that Tom has men-
ded his shirt is false, that Tom has mended his shirt, if true, would imply
that Tom is mending his shirt is false, and so forth for all implications in
[M2]. The very same sentence denotes an activity, however, just in case
we understand it as implying that Tom has mended his shirt is true (e.g.
in the sense of: Tom has done some ‘shirt-mending’), that Tom has been
mending his shirt is true, and so forth for all implications in [M3]. Occur-
rence types, this is the core intuition underlying the suggested definition,
are ways of ‘conceptually packaging’ an occurrence. Just as we ‘pack-
age’ the lexical content of a predication differently by combining it either
with an imperfective aspect or a perfective aspect, so we can ‘package’
a content also with a more complex processual information consisting of
a network of aspectual implications. On this approach, occurrence types
are denoted by complex verbal aspects. Since verbal aspects in general are
processual information about a mode of occurrence (e.g. the progressive
presents the occurrence as unbounded and ongoing, the perfective as a
bounded unit), also complex verbal aspects such as the ‘activity network’
or ‘accomplishment network’ denote modes of occurrence. In sum, on the
modal interpretation of occurrence types that I am suggesting here, the
answer to the question ‘what is an activity?’ runs as follows. Activities
are not a certain class of occurrences but they are modes of occurrences.
Occurrences are not fixedly linked to any of the ‘dynamic shapes’ that the
Vendler classification identifies. Occasionally we treat typical Vendlerian
‘accomplishments’, ‘achievements’, or ‘states’ as activities, and vice versa;
treating an occurrence as activity, accomplishment etc. means to concep-
tualize the occurrence as something that happens in a certain way. Our
concept of an activity is the concept of how something is happening – it
consists in the set of inferences specified in [M3i–iv] which we can draw
when we know that something has happened in this way. To reformulate
the conceptual knowledge encoded in [M3i–iv], if an occurrence is an
activity, then it occurs in a mode of occurrence that is characterized by
the following four features:
(C5) Completeness condition: According to (M3-iv), from ‘N is V -ing’ we
can infer ‘N has V -ed’, i.e. activities are always already completed
once going on.
(C6) Resumability condition: According to (M3-i and ii), from ‘N has V -
ed’ we cannot conclude ‘N is still V -ing’ nor ‘N is no longer V -ing’,
i.e. activities can be suspended and resumed.26
32 JOHANNA SEIBT
(C7) Recurrence condition: According to (M3-iii), from ‘N V s’ we can

infer that there are multiple recurrences of the same activity, i.e.
activities can recur.
(C8) Dynamicity condition: According to (M3-iv), from ‘N is V -ing’ we
can infer ‘N has been V -ing’, i.e. any present going on of an activity
is the outcome of its past going on.
2. FREE PROCESSES : A NEW CATEGORY
If an occurrence occurs as activity, then it is conceived of as an entity

that fulfills conditions (C5) through (C8) which correspond to the cat-
egorial inferences licensed by activity sentences. If free processes are to
be the truth-makers of (among others) sentences about activities, then the
category features of free processes must entail [M3i–iv] or (C5) through
(C8). But which category features will accomplish this task? Consider the
following two mereological features:
[1] Likepartedness: An entity of kind K is likeparted iff some of its spatial
or temporal parts are of kind K.
[2] Strict likepartedness: An entity of kind K is likeparted iff all of its
spatial or temporal parts are of kind K.
If we take the completeness condition (C5) to state that for most times t
during N’s V -ing it holds that N has V -ed at t, then the truth-maker (C5)
is compatible with both of these conditions. On the other hand, if we read
(C5) as saying that for all times during N’s V -ing it holds that N has V -ed,
then only strictly likeparted or strictly self-contained entities can qualify as
truth-makers for sentences about activities. Vendler took ‘activities’ to be
strictly likeparted (cf. C3 above); others criticized this decision because,
strictly speaking, in every occurrence – even a running, reading, sliding,
or seeing of something – we can identify phases: every occurrence can
be shown to contain change.27 Indeed, we can easily argue that not every
part of running is a running, just as not every part of water is again water
– the homeomerity of ’activities’ has been compared to the homeomerity
of stuffs and taken to hold only for a certain ‘grain-size.’ In my view this
discussion is partly misguided; when we discern phases in an occurrence
we are no longer conceiving of it as an activity but view one part of it as an
accomplishment – all that the discussion displays is possible ‘type shift.’
Nevertheless, the debate about the ‘granularity’ of ‘activities’ can be used
to motivate the following generalization of the predicate of homeomerity:
(HOM) Homeomerity: An n-dimensional (1 ≤ n ≤ 4) entity E of kind K

is homeomerous in dimension n iff (a) all or (b) some or (c) none of E’s
parts in dimension n are of kind K. In case (a) E is maximally likeparted,

in case (b) likeparted, and in case (c) minimally likeparted.28
An entity E that is homeomerous in this general sense can be the truth-

maker of sentences implying (C5) (in case E is maximally likeparted),
or of sentences about accomplishments (in case E is minimally likepar-
ted), and of sentences about sequences of accomplishments (in case E is
likeparted).
An entity that is homeomerous in the sense of (HOM) can be the truth-
maker of sentences about any extended occurrence. But this is not all. As
already intimated above and frequently observed in linguistic semantics,
there are striking inferential symmetries between sentences about activities
and sentences about stuffs or ‘masses’, and correspondingly between sen-
tences about accomplishments and sentences about things. Homogenous
masses (such as gold or water) are maximally likeparted in space, het-
erogenous masses or mixtures (such as fruit salad, furniture, or H2 O) are
likeparted in space, and things or ‘countables’ are minimally likeparted in
space. Where ontologists have taken note of these inferential symmetries,
they have failed to observe, however, that the denotations of nouns display
a similar flexibility in category implications as the denotations of verbs.
Just as the distinction between activity sentences and e.g. accomplishment
sentences is best considered a matter of complex verbal aspects (repres-
ented by networks of aspectual implications), so the distinction between
‘mass readings’ and ‘count readings’ of noun phrases is determined by
‘nominal aspects’.29 Most nouns in English are lexicalized either with the
category implications for masses (‘water’) or the category implications
for countable items (‘book’). But again, as in the case of verb phrases,
these lexicalized meanings specify merely defaults readings; supported by
morphosyntactic nominal aspect markers and, in particular, the context of
interpretation, speakers may shift the categorial implications from mass
to count or vice versa. In classifier languages where nouns are lexically
underspecified with respect to their categorial implications such nominal
type shifts are the rule – different classifiers ‘package’ nominal content in
different ways, e.g. as denoting a mass or as denoting a countable item.
But the phenomenon is also frequent in English; for instance, relative
to sentence context and utterance context, we can read chicken, car, or
grapes with the category implications for countable individuals or with the
category implications for a certain (functional) stuff (cf. I’d like some of the
chicken, get more car for your money, a pound of grapes contains about 50
grapes). Even proper nouns or pronouns may be given a ‘mass’ interpret-
ation – you may wish that someone were less George and more president,
34 JOHANNA SEIBT
or join the American army to “be all you can be”. In short, the inferential
symmetries observed between activities and masses, or accomplishments
and things, respectively, suggest that there are two types of conceptual
‘packaging’: contents of both verbs and nouns may be packaged with the
category implications of non-countable items or with the category implic-
ations of countable items. Non-countable items are maximally likeparted
or likeparted, countable items are minimally likeparted.
But free processes are not merely homeomerous entities, they fulfill
an even stronger mereological predicate. Let us contrast likepartedness as
above in [1] with self-containment, and homeomerity as in (HOM) with
automerity:
[3] Self-containment: An entity E is self-contained iff the spatiotemporal
region in which all of E occurs has some spatial or temporal parts in
which E (i.e., all of E) occurs.
(AUT) Automerity: An n-dimensional (1 ≤ n ≤ 4) entity E of kind K

is automerous in dimension n iff (a) all or (b) some or (c) none of E’s
parts in dimension n are an occurrence of E. In case (a) E is maximally
self-contained in case (b) self-contained, and in case (c) minimally self-
contained.30
As witnessed by (C7) above, if we conceptualize an entity as an activity,

we conceptualize it as a recurrent entity, and the same holds for masses: if
swimming laps is taken to occur as an activity, and if ‘grapes’ is taken to
specify a stuff, then swimming laps – the very same activity – and grapes
– the very same stuff – can both occur now and then, or here and there at
the same time. The feature of self-containment is one way to articulate in
mereological terms an ancient ontological intuition, namely, that there are
entities which occur multiply as the numerically (not just qualitatively!)
same entity. Self-contained entities are obviously not particular entities, i.e.
entities that occur in a unique spatiotemporal location at any one time, and
as long as one adheres to the presuppositions of the substance-paradigm
one is forced to class an entity that recurs as the numerically same entity
as a universal.31 At first glance this seems smooth enough: universals are
said to occur multiply in space, their identity conditions do not involve
its spatiotemporal location(s), and they are not countable by means of
spatiotemporal extent. But in order to occur spatiotemporally, traditional
universals, whether abstract or concrete, always depend on a particular
that is the logical subject of their qualitative determination. In contrast, we
may well think self-contained entities as multiply occurrent without being
‘attributed’ to any particular – you and I, we may have the same wine in
our glasses and when we drink from them we engage in the same activity,
but there is no ‘logical subject’, not even the spacetime region, to which
the wine or the drinking need to be attributed to.
Moreover, a self-contained entity can well be considered an individual,
i.e. something we are able to refer to. What we refer to must be ‘re-
referrable-to’, and this requires, vide Strawson, transtemporal sameness.
But, pace Strawson, referential reidentification does not imply particular-
ity as required by spatial co-ordinatization. Co-ordinatization in arbitrary
parameter spaces or functional individuation is all that is needed. Ana-
phoric references as in yesterday I saw in Tom’s sailboat the wood we
used to have in the kitchen or That sport has been practiced since the
middle ages indicate that self-contained entities easily qualify as individu-
als. Functionally individuated entities are determinable entities, at least
with respect to their spatio-temporal location; once we admit that self-
contained entities are individuals, we thus break away from what might
be considered the core presupposition of the myth of substance: that all
individuals are particular and fully determinate entities.
Altogether, then, free processes are concrete, individual, automerous
entities; since they are automerous, they are by implication functionally
individuated and determinable. While traditional ontology has priorit-
ized countable particulars (things), in free process theory countability
and particularity are merely the limiting cases of non-countability (self-
containment) and determinability. Some free processes are minimally
self-contained – things and developments (taken as functional wholes) do
not occupy any proper part of the spatial region they occupy, and that
is all there is to countability. Some free processes have a functionally
individuating description which happens to include the specification of a
unique spatiotemporal location (e.g., snowing-at-t-in-location-s), and that
is all there is to particularity. But note that particularity as such does not
imply full determinateness in the sense of a Leibnizian infima species.
Free processes are determinable ‘functional stuffs’ even when occurring
as particulars: as detailed below the amount of a free process α is itself a
free process β which has determinate spatiotemporal location but might be
indeterminate in other respects.
To my knowledge all extant ontological schemes of process ontologies
conceive of processes as concrete particulars, that is, as filled space-
time regions or the discrete particular fillings of spacetime regions.32 The
paradigm example of a process is taken to be either the particular perform-
ance of a human activity, such as a particular running or reading, or, the
particular occurrence of a ‘subjectless’, ‘absolute’, or ‘pure process’ (C.D.
Broad; W. Sellars): a snowing or thundering in a particular spatiotemporal
36 JOHANNA SEIBT
region. In contrast, free processes are not only ‘free’ in the sense that they
are not alterations in a subject, they are also free in the sense that they are
not ‘bound to’ a specific spatiotemporal region. Intuitively speaking, free
processes are goings-on as expressed by ‘feature-placing statements’ of a
more or less ‘placing’ sort: it is raining now, it is itching here, there’s good
sailing all along the coast, a photon is traveling from the sun to the earth.
So far I have mentioned category features of free processes that are
implied by or compatible with conditions (C5) and (C7) above, but I
have not commented on the category feature ‘dynamicity’ which dove-
tails with conditions (C6) and, in particular, (C8) above. Elsewhere I
argue, based on an analysis of Aristotle’s notion of energeia, the process-
ontological core thesis, namely, that being and going on or dynamicity
as ‘self-production’ are co-intensional concepts.33 The notion of dynam-
icity supplied there distinguishes free processes from the static expanses
of ‘four-dimensionalism’, and allows us to stipulate free processes even
as truth-makers for many Vendler ‘states’, reading predications such as
the ball’s being red and circular as though they were to involve the
progressive.
The task of Free Process Theory is to show that the notion of a free pro-
cess is wide enough to accommodate the inferential roles of a large number
of classificatory terms (thing, event, action etc.); that is, to show that the
truth-makers of English and its translation equivalents consist of nothing
else but free processes: goings-on, automerous in this or that dimension,
simple and complex, slow and fast, evenly and with culmination, creating
(inferential stand-in’s for) particular τoδε τι’s ‘on the go’.
3. FPT: A MEREOLOGY ON FREE PROCESSES34
Classical Extensional Mereology (CEM) is geared to the reading of the

part-whole relationship that applies if the parts and wholes in question
are geometrical regions. But, as has been noted in the literature, in ap-
plication to other ‘arguments’ the relation axiomatized in CEM does not
capture equally well our intuitive usage of ‘part-whole’ concepts. Mostly
this deficiency of CEM is attributed to a failure of the extensional identity
principle of CEM (Proper Parts Principle).35 In contrast I suggest dropping
the transitivity axiom.36 The relation modeled by classical mereology (is
a piece of, is an extension part of) is not the basic sense of the ‘is part
of’ relation which has functional overtones that disturb the transitivity
of the relation.37 (The fact that the transitive ’is a piece of’ relation has
taken centerstage in the theoretical modeling of parthood reflects nicely
the particularist bias of the substance paradigm.) Rather, the role of ar-
ticulating the fundamental sense of parthood should better be assigned to

the non-transitive part-relation holding among stuffs and activities – ‘is
part of’ (belongs to, comes with, is involved in) is more basic than ‘is a
(spatiotemporal) part of.’ In order to contrast ‘is part of’ in this widely
functional, non-spatiotemporal sense from the classical geometric reading,
I refer to it as the ‘common sense part relation.’ But note that the non-
transitive common sense part relation can be used to define the familiar
transitive relation of parthood as holding between geometric regions or
‘extensions’ (‘extent part’, abbreviated as ‘<ext ’ and diversified into ‘<temp
and ‘<sp ’).38
The common sense part relation (read: ‘is part of’; abbreviated ‘ ’)
of FPT is defined on the field of free processes as an asymmetric and
irreflexive relation:
(Ax1) x y → ¬y x
(Ax2) ¬x x
Parthood-or-identity is then defined as
(D1) x y↔x y ∨ x = y.
In FPT parthood as expressed by ‘ ’ reaches only into the immediate or

‘first-level’ parts of a process. To express part relations with ancestors
a notion of n-part is defined (which corresponds, for variable n, to the
transitive closure of ‘ ’; abbreviated by ‘n ’):
(D2) ∀n ∈ N; if n = 1 then x n y ↔ x y; if n ≥ 2 then x n y ↔

∃z (x z & z n−1 y)
The expression ‘+m-part’ refers to all n-parts with m ≥ n. FPT is ‘ex-

tensional’ in the sense that the Proper Parts Principle (PPP) supplies the
identity principle for free processes, but note that it is defined in terms of
an n-part, i.e. only as axiom schema:39
(Ax3) ∀z(z n z↔z n y) ↔ x = y.
Since ‘ ’ is not the relation of extent part, the well-known counter-

examples to the (P P P ) (e.g. undesired identifications of functional units
with material constituents) are avoided. The relation of ‘overlap’ can be
defined as usual,
(D3) x ◦ y ↔ ∃z (z x&z y),

38 JOHANNA SEIBT
but due to the non-transitivity of ‘ ’ the overlap of two entities we receive

two types of non-overlap, discreteness (D4) and disjointness (D5):

(D4) x y ↔ ¬∃z (z x & z y)
(D5) x | y ↔ ¬∃z (z n x & z n y)
The supplementation principle in FPT is defined in terms of discreteness

rather than disjointness:

(A4) x y → ∃z(z x & z y)
This allows for wholes whose parts cannot be completely ‘disentangled’,

such as, for example, activities performed simultaneously by a living or-
ganism (walking and jumping puddles) which each involve the same com-
ponent activity (lifting the left leg). Similarly, due to the nontransitivity of
‘ ’ product and sum, defined in the usual way:
(D6) prod(x, y) =df ιz(∀w(w x&w y) ↔ w z)
(D7) sum(x, y) =df ιz(∀w(w ◦ z → w ◦ x ∨ w ◦ y))
deviate to some extent from their classical meanings.40 But product and
sum are associative and commutative.41 The inverse of the sum of free
processes, their difference, is defined as:

(D8) dif(x, y) =df ιz(z x & z y).
The sum of two free processes is itself a free process with arbitrarily
scattered spatiotemporal parts; no causal interaction is implied. For in-
stance, walking and chewing gum is a sum of free processes, but so is
a phone conversation, the French revolution, or inflation. Since the sum
operation does not ‘reach’ below the level of 1-parts, we can distinguish a
large number of different process combinations in terms of additional con-
ditions on the structure of +2-parts (see section 4.2 below). Some process
combinations ‘create’ +2-parts, others ‘cancel’ existing ones (phenomena
of ‘emergence’ and ‘suppression’). The general term used in FPT for all
process combinations which ‘effect’ a change in +2-parts is ‘interference’
(abbreviated by ‘I (x, y)’). To highlight that the interference of two free
processes α and β is itself a free process γ , read ‘γ = I (α, β)’ as ‘γ is the
interfering of α and β’.
The field of ‘ ’ consists of more and less specific free processes
(e.g., lifting the right leg running, nodding one’s head greeting,
photosynthesis plant growth). Among these is spacetime (ρ0 ) and its spe-
cifications representing different, more or less determinate spatiotemporal
regions or ‘spacetimings’ (ρi ).42 If process α is located in a determinate
region, then there is a minimally self-contained process β which is the
interfering of α and ρ; β is called an amount of α, abbreviated ‘[α]’:
(D9a) ∀x = ρi : [x] =df ιz (∃ρi (z = I (x, ρi ))
Amounts of processes also ‘carve out’ amounts of spacetime:
(D9b) ∀ρi : [ρi ] =df ιz (∃x (z = I (x, ρi ) & ¬(ρ0 x))43
Amounts of processes have a determinate location – they are located in
(i.e. they are inferences with) an ultimately specific space-time region.44
In contrast, quantities of processes are located indeterminately – they are
interferences with a less specific spatio-temporal region (in town, in the
accelerator).45 For all processes it holds that that they interfere with ρ0 , i.e.
that they occur somewhere in spacetime. This is the occurrence axiom:46
(Ax5) ∀x[∃y(x y∨y x) → ∃z, ρi (z = ((x, ρi ))]
That a free process occurs in a certain spatiotemporal region (for in-
stance, that partying is going on at Roger’s house tonight) means in FPT
that two processes α and β (partying, being-the-extension-of-Roger’s-
house-tonight) are superposed to form an amount of each; their sum is
an interfering, i.e. a third complex process γ , which overlaps with both
partying and being-the-extension-of-Roger’s house-tonight) in its 1-parts
but differs in its +2-parts from the +2-parts of either of them. For instance,
assuming that the 1-parts of partying are communicating and human group
presence, we might find that joking and discussing literature are part of
the communicating that goes on at Roger’s house tonight, but these are not
+2-parts of communicating or human group presence. The idea behind
definitions (D9a) and (D9b) is thus simply this: the occurring of a free
process α in a region ρi is a process β which is a more specific version
of α as well as of ρi . It does not hold, however, that every specification
of a free process α results in an amount of α. To express the connection
between amount and specification of a free process more clearly, we first
need to look at the FPT-definition of extent-part, i.e. the basic relation of
Classical Extensional Mereology:
(D10) [x] ≤ [y] ↔ ∀ρi (I (x, ρi ) → I (y, ρi ))
The transitivity of ‘≤’ is warranted by
(Ax6) [x] ≤ [y] & [y] ≤ [z] → [x] ≤ [z].
40 JOHANNA SEIBT
On the basis of ‘≤’ the FPT-predicate ‘x is a phase in y’ can be introduced;

a phase of α is a part of α an amount of which occurs in every amount of
α:
(D11) phase(x, y) ↔ x y & ∀z(z = [y] → ∃w (w = [x] & w < z)).
A free process α with phases can then be said to be a specification of a

process β iff every phase of α is an amount of β:
(D12∗ ) specif(x, y) ↔ ∀z(phase(z, x) → z = [y])
For example, since all the phases performed in cooking lasagne are
amounts of cooking, cooking lasagne is a specification of cooking. For
processes without phases, such as playing guitar or being a poodle, the
quantification is over the spatiotemporal parts of any amount of the specific
process. So the full version of the FPT-definition of specification runs as
follows:
(D12) specif(x, y) ↔ {∀z(phase(z, x) → z = [y]) ∨ ∀z, x(z ≤ [x] →
z = [y])}
If Max is a poodle, wherever there is an amount of being-Max, there is

an amount of being-a-poodle. An ultimate specification of a process is a
process for which there is no further specification within the domain of
FPT.
(D13) ult-specif(x, y) ↔ specif(x, y) & ¬∃z(specif(z, x))
It is crucial to realize that in FPT an amount of a process α is a specification

of α, but not necessarily an ultimate specification of α. On the other hand, a
process α that is an ultimate specification of a process β must be specified
with respect to spatiotemporal occurrence, i.e. it must be an amount of [β].
This follows from definitions (D9b), (D12), (D13), and the occurrence
axiom (A5). Ultimate specificity implies not only spatiotemporal occur-
rence but also ultimate specificity of spatio-temporal occurrence. Thus we
can state the following Particularity Theorem, which says that ultimate
specifications can fulfill the inferential role of traditional particulars, since
they have an ultimately specific and thus unique spatiotemporal location:
(T1) ∀x [∃y ult-specif(x, y) → ∃z, ρi (x = I (z, ρi ) &

¬∃ρj , v (specif(ρj , ρi ) & v = I (z, ρj )))]47
A third basic relationship between processes, besides phase and spe-

cification, is that of a stage of a process. When we say that applying the
primer is part of varnishing the cabinet, that tadpoles become frogs, or that
being an adolescent is part of being human, we imply that some but not all
of the classificatory predicates that characterize the whole development,
apply also to a spatiotemporal part of it. In FPT this is expressed as the
requirement that for a process amount [α] and a spatiotemporal part [β]
of [α], [β] and [α] have some but not all specification relationships in
common:
(D14) stage(x, y) ↔ ∀z(z = [x] → ∃w(w = [y] & z < w)

& ∃z (specif(x, z) & specif(y, z)) & ∃z, v (specif(x, z) ↔
¬specif(y, z) & ¬specif(x, v) ↔ specif(y, v))
Axiom (A5) ensures that (D14) is not vacuously fulfilled. Since according
to (D14) anything that has stages is ‘traceable under some sortal’ (i.e.,
in the FPT idiom: is a specification of some process), a dynamic entity
is something that can be viewed as both changing and transtemporally
identical.48
All entities in the field of ‘ ’ are dynamic, as postulated in the Process
Axiom:
(Ax7) ∀x∃y(x y∨y x) → dyn(x)].49
Axioms (Ax5) and (Ax7) in combination imply that all free processes have
spatiotemporal parts, i.e., the extent-part relation ‘<’ has no atoms in FPT.
4. TOWARDS A TYPOLOGY OF PROCESSES
4.1. Classification by automerity pattern

Free processes can be coarsely classified into five types, according to cer-
tain characteristic mereological conditions. Recall the distinction between
three types of automerity, maximal self-containment, self-containment,
and minimal self-containment (cf. (AUT) above). In terms of these three
predicates various automerity patterns can be differentiated and associated
with basic process types.50
(1) Type 1 processes are temporally maximally self-contained and spa-

tially unmarked; their prime examples are activities, e.g., a running,
raining, or reading taken to occur in the activity mode. The complete-
ness condition (C5) above requires for activities maximal temporal self-
containment; but spatially there may be maximal self-containment (fall-
ing), self-containment (raining), or minimal self-containment (Tom and
42 JOHANNA SEIBT
Kim’s pair-dancing). In the formal idiom of FPT the characteristic auto-

merity condition for type 1 processes can be restated as follows. (Recall
that ‘I (α, ρi )’ stands for ‘an amount of α occurs in spatiotemporal region
ρi ’).
(D16) α is type 1 iff ∀x(x <temp I (α, ρi ) → x = I (α, ρj ).
(2) Type 2 processes are temporally minimally self-contained while their

spatial self-containment is unmarked; prime examples of such processes
are occurrences in the accomplishment mode (developments, events). The
performance of Mozart’s Clarinet Concerto, for instance, is temporally
minimally self-contained (no part of the performance of the Concerto
is itself a performance of the Concerto) and spatially minimally self-
contained; other events, however, like the communal singing of the na-
tional anthem, are equally temporally minimally self-contained but do have
spatial parts that fulfill the description of the whole. To restate:
(D17) α is type 2 iff ∀x(x <temp I (α, ρi ) → ∼ (x = I (α, ρj ))).
(3) Type 3 processes are spatially minimally self-contained and temporally

maximally self-contained; their prime examples are entities conceived of as
things and discrete expanses of matter. According to our common ways of
talking, a table or piece of rock do not take time but ‘endure’, i.e. every
temporal part of them fulfills the description we apply to the temporal
whole. To restate:
(D18) α is type 3 iff ∀x(x <temp I (α, ρi ) → x = I (α, ρj )) &
∀x(x <sp I (α, ρi ) → ∼ (x = I (α, ρj )).
(4) Type 4 processes are spatially and temporally self-contained. Their

prime examples are oscillatory sequences of developments as well as
heaps, e.g. running conceived of as a sequence of repetitive movements
or water taken as assembly of molecules. When authors argue for the
limited homogeneity of activities or masses, then they actually speak
about oscillatory sequences and heaps, assigning to an occurrence like
running a phase that is not a running (lifting the left foot), or to a heap
(coarse-grained mixture) like H2 O a spatial part (oxygen molecule) that
is not H2 O. Note, however, an important asymmetry between oscillatory
sequences and mixtures: the latter are always also temporally maxim-
ally self-contained, since they are taken to ‘endure’ like things. Formally
restated:
(D19) α is type 4 iff ∃x (x <temp I (α, ρi ) & x = I (α, ρj )) & ∃x(x <sp
I (α, ρi ) & x = I (α, ρj )).
(5) Type 5 processes are spatially and temporally maximally self-

contained. Prime examples of such processes are entities conceived of
as masses or stuffs proper. Once water is not viewed as a mixture but
as stuff, it is spatially maximally self-contained, and, since it endures, it
is temporally maximally self-contained. Taking a ‘substantivist’ view on
space and time (as this is done in the current version of FPT), space, time,
and spacetime are also type 5 processes; further candidates are perhaps
secondary qualities such as colors or sounds.
(20) α is type 5 iff ∀x (x <temp I (α, ρi ) → x = I (α, ρj )) & ∀x

(x <sp I (α, ρi ) → x = I (α, ρj )).
It is important to note, however, that the conditions stated in (D16) through

(D20) figure only as necessary conditions in the FPT-interpretation of
common sense classificatory predicates such as ‘activity’, ‘event’, ‘thing’,
‘stuff’, and ‘quality’. The FPT-definition of thinghood, for example, re-
quires type 3 processes to be spatially unified, transportable, and to possess
functional shape, which distinguishes them from discrete expanses of
matter.51
4.2. Classification by dynamic parameter analysis

A more fine-grained classification of free processes can be achieved within
a four-dimensional parameter space based on the following evaluative di-
mensions: (a) participant structure (types and roles of participants); (b)
dynamic composition; (c) dynamic shape; (d) dynamic context.
4.2.1. Participant structure

From an Aristotelian point of view, talking involves one substance, arguing
two, and the ‘elemental transformation’ snowing none; whether chopping
involves more than one substance remains unclear, due to the inclusive
status of composite artifacts. Such differences in the participant structure,
i.e. the types and roles of participants in occurrences, are often encoded
linguistically: partly in the explicit argument structure of verbs (transitive
vs. intransitive), partly they are expressed by prepositions (compare hit
the door vs. hit at the door), partly they are signaled by special verb
forms (e.g., passive, accusative, ergative forms).52 In general, however,
it is not possible to use syntax as a reliable indicator of the participant
structure. Many sentences with intransitive verbs do not involve exactly
one agent. Some involve implicit references to locations, times, or ob-
servers – cf. ‘he disappeared’, ‘the wedding took place’, ‘the game was
disappointing’ – others express qualifications of dynamic shape or dy-
namic context of a process, such as ‘the rain increased’, ‘the rate of change
44 JOHANNA SEIBT
remained constant’, ‘this development could not be stopped’, etc. Some-

times only context can determine the participant structure of a sentence
with an intransitive verb: ‘Fred stumbled.’ The framework of FPT allows
us to express rather subtle differences in participant structure. Compare for
instance the following six distinctions:
(1) A process α is a one-agent process iff there is an amount of α that is

part of a temporal part of exactly one process γ of type 3 or 4 (thing, piece
of matter, or stuff), and no part of α is part of a temporal part of another
process of this type at that time.53 For example, walking or rolling are one-
agent processes insofar as there are amounts of walking or rolling that are
a part of a temporal part of Fred or this ball, respectively, and of no other
thing (piece of matter or stuff) at that time.
(D21) one-agent process (α) iff ∃[α], [β], x, ∃1 γ3,4 ([α] [β] & x =
[γ ] & [β] ≤temp [γ ] & ∀δ3,4 ([β] <temp [δ] → δ = γ ))54
(2) A process α is a one-agent-one-patient process iff there is an amount

of α that is part of a temporal part of process γ , and there is a stage δ of α
which interferes∗ a temporal part of a process ε (where γ and ε are not spe-
cifications of α). For example, touching, picking up, looking at, receiving
are processes with stages that are part of the temporal parts of the ‘patient’
involved. The set of agents and patients is here unrestricted as far as type
is concerned – activities, events, things, stuffs, and phenomenal properties
can shatter, decorate, disturb, overshadow, or accompany activities, events,
things, stuffs, and phenomenal properties.55
(D22) one-agent-one-patient process(α) iff ∃[α], [β], [δ], [φ], x, y,

∃1 γ , ε([α]) [β] & x = [γ ] & [β] ≤temp [γ ] & stage(δ, α)
& [δ] [φ] & y = [ε] & [φ] ≤temp [ε & ¬(specify(γ , α) ∨
specif(ε, α))).
(3) Since more than one agent may collectively or separately affect
one or more patients collectively or separately, we can diversify (D22)
further into many types of (collective)-n-agent-(collective)-m-patient pro-
cesses: (a) collective n-agent m-patient processes (as expressed in the
sentence ‘Between the two of them, Max and Tom ate three pizzas’) (b)
n-agent-collective-m-patient processes (as in: ‘Max and Kim each carried
three pizzas’), (c) collective-n-agent-collective-m-patient processes (as in:
‘Max and Kim are assembling the parts’) and (d) n-agent-m-patient pro-
cesses (as in: ‘While Max played guitar, Kim read a book’). Collective
agency/patiency is in FPT expressed in terms of process interference and

separate agency/patiency in terms of process sums.56
(4) Collective agency and patiency may further be differentiated according

to the the degree of entanglement of agents and patients involved. Collect-
ive agency (patiency) as defined in terms of interference expresses merely
the type of togetherness resulting from spatiotemporal co-occurrence. But
in many cases co-occurrence is accompanied by dependence relationships
holding among the involvements of the agents or patients. In (a) concerted
(or weakly entangled) n-agency (as in: ‘Max and Tom lifted the piano’
or ‘sticklebacks perform a complicated courting ritual’) or concerted (or
weakly entangled) n-patiency (as in: ‘Max drank the gin-and-tonic’, ‘the
measurement showed an anti-correlation of the electron’s spin and the
positron’s spin’) the agent/patient process is an interference (weakly emer-
gent product, see below) of its 1-parts (component actions/passions such
as ‘Max’s pulling the piano up’, ‘the male stickleback’s zig-zag dance’, or
‘the electron’s spin being measured’, ‘the gin’s being drunk’, resp.) which
can be performed independently. In contrast, in (b) reciprocal (or deeply
entangled) n-agency/patiency (as in: ‘Max and Kim discussed the issue’,
or ‘the roof is supported by intricate timbering’, or, for patiency: ‘measure-
ment M effects a superposition of states with spin 1/2’), the interference
concerns not only the 1-parts of the composite action. Here the 1-parts
(e.g., ‘Max’s discussing the issue’, ‘the ridge beam’s lying horizontally
above the ground’, ‘spin 1/2’) are themselves interferences of a certain
kind and cannot occur independently of what they are interferings of.
They are either weakly emergent products as just introduced, or strongly
emergent products, i.e., processes that are interferings of a collection of
processes which each contain at least one other member of the collection
as part (more on weak and strong emergence below).
(5) Finally, participant structures are classed according to the types and
identities of agents and patients involved, which allows, for example, for
useful distinctions between: (a) bodily movements, (b) reflexes and afflic-
tions, (c) undertakings, i.e. deliberate bodily movements such as push-ups,
where a human agent is acting on herself ‘qua other’ (i.e. human body or
mind), and (d) basic actions, where the agent is acting on herself ‘qua self.’
4.2.2. Dynamic composition

The evaluative dimension called dynamic composition addresses dynamic
phenomena discussed under headings such as ‘generation’, ‘destruction’,
‘emergence’, ‘downward causation’, ‘self-organisation’, and ‘complexity.’
46 JOHANNA SEIBT
Just as in classical mereology, in FPT the sum γ of two items α and β is

the item γ that is overlapped by anything which overlaps α or β. However,
due to the non-transitivity of the part-relation in FPT this definition leaves
room for further differentiations in ‘additivity types.’
(Add-1) A process γ is simply additive iff γ = sum(αi ), and

there are no interferences∗ on the n-parts of the αi , i.e. the +2-
parts of γ are the +2-parts of all αi .
Intuitively, a simple addition of processes consists of two non-interacting

processes, such as my walking in Texas and radioactivity on Jupiter, where
(presumably) none of the constituent processes of my walking in Texas or
of radioactivity on Jupiter cancel each other out by interference, nor any
new processes are generated.
(Add-2) A process γ is linear additive iff γ = sum(αi ) and for

any +n-part β of γ , β there is a despecification of β which is
also a despecification of some +n-part of every αn , 1 ≤ n ≤ i.
Linear additivity generates complex processes which, intuitively speak-

ing, have the same kind of features as their components but different
magnitudes, the prime example being the vector addition of forces.
(Add-3) A process γ is weakly emergent on processes αi iff γ

is a +2-part of β = sum(αi ) and γ is not a +2-part of any αi
and no n-part of γ is also an n-part of some αi .
A weakly emergent process has functional features that none of its com-
ponent processes has. A wheel has the weakly emergent feature of being
circular which none of its components has, a chord has a harmonical role
which none of its component sounds has, a simple feedback cycle has a
control function which none of the components have.
(Add-4) A process γ is strongly emergent on processes αi iff

γ is weakly emergent on processes βi which are each weakly
emergent αi .
The wheel’s rolling for example is strongly emergent on the structural

components of the wheel and of the supporting surface, since it is weakly
emergent on the wheel’s circularity and on the inclination of the surface;
self-organizing systems such as biological organisms and economies are
strongly emergent on the specific physiological and economical processes
occurring here and now since they are weakly emergent on hierarchies of
functional units whose levels are related by weak emergence.57
Non-linear addivity can take many forms. On the one hand, there are
many ways to combine and reiterate weak and strong emergence. For
example, ‘musical processes’ in E. Christensen’s sense (cf. his contribu-
tion to this volume) are, I believe, an illustration of processes that are
strongly emergent on musical units (phrases, harmonies) which themselves
are strongly emergent on tone sequences. On the other hand, we can distin-
guish different types of non-linear dynamic composition according to the
dependence structures among the n-parts of the emergent process γ . For
example, consider Bickhard/W. Christensen’s notion of an ‘autonomous
system’.58 An autonomous system – a ‘generalization of the concept of
autocatalysis’ – is characterized by a very specific form of organizational
interdependence of its structural components: it “interactively generates
the conditions required for its existence”.59 This involves interdependence
of the system’s components at different degrees of directness:
(D23) Process γ is an autonomous system iff γ is strongly emergent

on a set of processes αi and for any 1-part β of γ there is an
n-part α of γ such that β α and α β.
In FPT with its non-transitive part-relation we can – unlike in systems of

classical mereology – painlessly express the fact that an organism’s (= γ )
blood flow (= α), cellular metabolism (= β), and motor activity (= δ)
mutually condition each other: (α, β δ) & (β, δ α) & (α, δ β),
turning γ into an autonomous system.
4.2.3. Dynamic shape and dynamic context

The details of the third and the fourth dimensions of evaluative parameters
in FPT are still very much under construction. ‘Dynamic shape’ relates
to differences with respect to the ‘course’ or flow of a process, as these
are expressed, for instance, by verbal aspects. Most verbal aspects (e.g.,
perfective, progressive, repetitive, ingressive, egressive) correlate naturally
with simple shapes of trajectories in phase space, or parts of such traject-
ories and thus admit of mereo-topological definitions on phase space. But
processes may also be differentiated with respect to how they affect their
dynamic context. The dynamic context of a complex process may differ
with respect to (a) linear causal consequences, (b) non-linear causal con-
sequences in their immediate dynamic environment, such as disturbances
in the air flow around a kite, and those with (c) non-linear consequences in
the generative environment of the process, such as changes in ecosystems
that alter selection pressures.60
48 JOHANNA SEIBT
5. C ONCLUSION
I have sketched here some elements of the heuristics and construction of

Free Process Theory, a monocategoreal process-ontological scheme which
operates with dynamic, multiply occurrent individuals. Abandoning the
traditional ontological bias in favor of countable particulars, FPT situates
all entities on a gradient scale of determinability and automerity where
particularity and countability merely present the limiting cases (minimal
determinability and minimal automerity, respectively). FPT was originally
developed in 1990 in the course of an attempt to expose and undercut
substance-ontological presuppositions that hamper satisfactory solutions
to the classical problems of individuation, universals, and persistence.
Later studies work out FPT’s built-in (dis)solutions of these three classical
problems,61 develop an account of thinghood, and explore the scheme’s
possible feasibility for the interpretation of quantum-physical entities.62
My primary goal here was to draw attention to two facts. First, as our
reasoning about activities displays, we are agentively well-familiar with
instances of ‘free processes’, a new ontological category. Second, a non-
transitive part-relation between free processes can be used to define various
specific part-relations (extent part, functional part, material part, con-
struction part etc.), the spatiotemporal location of processes, distinctions
between common sense ‘categories’ such as thing, event, stuff etc., and
many different types of fine-grained constituent relationships, including
those involving internal interdependencies, dynamic modifications, and
emergent complexity.
NOTES
1 Cf. Seibt 1990b, 1995, 1996a, 1996b, 1996c, 1999, 2000d, 2002, 2003.
2 Cf. Sellars 1981, Zemach 1970.
3 This at least is the view of the – currently predominant – Carnapian tradition in analytical
ontology. On historical and methodological issues cf. Seibt 1996d, 1997, 2000a, 2000c,
2002a, where I also explain in which sense ontologies are language-transcendent (i.e.,
applicable to the class of functional equivalents of L) – a point which is here omitted
for reasons of simplicity.
4 That is, for its basic types of entities (categories) the domain theory specifies certain
features (category features) with suitable explicit definitions. The inferences licensed by a
sentence S are justified if they can be shown to follow from (the definitions of) features
of categories that are part of the truth-maker of S. For example, consider the English
sentences:
(1) ‘My Ford is a rather old car’.
(2) ‘A car is a means of transportation’.
Sentence (1) licenses the inference:

(3) ‘What is left of my Ford is not my Ford’,
but (2) does not license the sentence analogous to (3):
(4) ‘What is left of a car is not a car’.
That (1) licenses (3) (while (2) fails to license (4)) has traditionally been explained by the
fact that the ontological correlate of ‘my Ford’ (but not of ‘a car’) is a primary substance,
which has the category feature of particularity or unique occurrence in space defined as
entailing (3).
5 This is one way to read the Carnapian postulate of foundedness in the Aufbau, cf. Seibt
2000a. For a more detailed description of the relationship between category and model
cf. 2000c. Note that the ‘model’ of an ontological category is denoted by an ultimate
genera term of L. One of the primary difficulties for theories of tropes (or ‘moments’)
consists, incidentally, in the fact that the category ‘trope’ lacks a model in this sense – in
English there is no ultimate genera term expressing the genus of ‘this red’ versus ‘that red.’
This lacuna is covered up by the tropist’s quick move to technical jargon like ‘property
instances’ or ‘exemplifications of attributes’, which does not found the category in the
required sense.
6 Compare the contribution of M.-L. Gill in this volume.
7 Ryle 1949: 149.
8 Cf. Vendler 1957, Kenny 1963.
9 Kenny 1963: 172.
10 Note that (C1) through (C5) are already used by Aristotle to distinguish energeia from
kinesis.
11 Ibid.
12 The critical interaction with Vendler’s and Kenny’s classification begins with Mourela-
tos’ seminal discussion from 1978. Cf. also Taylor 1985.
13 Different sentential contexts can change the ‘action type’ of the verb; compare: Tom
smoked a cigarette – Tom smoked cigarettes – Mary pushed the cart to the shop – Mary
pushed the cart for hours.
14 Spell-bound by the ‘logical grammar’ developed by native speakers of Germanic lan-
guages, analytical ontologists largely continue to overlook the inferential significance of
verbal aspects. This is not surprising since in Germanic languages there are few, if any,
verbal aspect markers; differences in ‘aspectual meaning’ are mostly expressed by peri-
phrasis. In Romance and Slavic languages aspectual meaning features more prominently;
compare the difference between the French Il regnait trente ans (he reigned for thirty
years, Imperfect, expressing the so-called imperfective aspect) and Il regna trente ans (he
had a reign for thirty years, Past Definite, expressing the so-called perfective aspect). As-
pectual meaning consists in information about the dynamic organization of an occurrence
or its relationship to other occurrences as backgrounding or incident, i.e. verbal aspects
characterize an occurrence as ongoing (progressive), attempted (conative), about to begin
(prestadial), just begun (ingressive), in the middle (continuous), about to end (egressive),
frequently recurring (iterative), habitual, or as a factual unit (perfective) etc. Cf. Comrie
1976 and Dik 1997.
15 In the linguistic literature occasionally both the shift of verbal aspects and its effect are
referred to as ‘verbal aspect shift.’
16 Cf. for example Dowty 1979.
50 JOHANNA SEIBT
17 Cf. Sasse 2002.

18 Cf. for example Dowty 1979, Bache 1995, Dik 1997, Smith 1997, Kearns 2000,
Rijksbaron 1989.
19 The exception being achievements; compare for instance Dik’s classification (1997),
to my knowledge the most diversified, which countenances 10 categories: situations
(which are either states or positions), events (which are, if processes, either dynamisms
or changes), and actions (which are either activities or accomplishments).
20 For example, a predication P is said to be telic iff it can be combined with ‘almost’
or ‘within n minutes (hours/days/years)’ or used in constructions with ‘finish’ or ‘stop.’
Altogether there are 9 linguistic criteria of telicity, which determine three slightly different
notions of telicity: mereological, topological, and modal telicity. For this and a detailed
discussion of extant linguistic definitions of dynamicity cf. Seibt, forthcoming, section 2.4.
In contrast, none of the definitions of dynamicity offered are successful: they are either
circular, or blur the distinctions between telic and atelic dynamicity, or between ‘static’ and
‘dynamic’ atelicity. In short, there is a strong intuitive difference between two classes of
situations – Vendler called them ‘states’ and ‘activities’ – which both fulfill the condition
of distributivity and homeomerity (cf. C3 and C4 above), but this difference, it appears,
cannot be defined at the lexical level alone without involving aspectual information.
21 Cf. for example Smith’s ‘two component theory’ (1997) and, in particular, so-called
‘selection theories of aspect’, as collected in Breu 2000.
22 Cf. Comrie 1976: 3 and 18; Dik 1997: 222, reformulating a comparison by A. Isačenko.
23 Cf. Seibt forthcoming.
24 In particular, selection theories tell us little about whether verbal aspect shift indeed
amounts to a conceptual type shift. All we know about the processual meaning of a sen-
tence such as Tom was crossing the street is that it denotes the phase of an occurrence with
phase and specified boundary; but it remains unclear whether our conception of such a
phase differs from the phase denoted by the sentence Tom was running, i.e., whether the go-
ing on of an ‘accomplishment’ and the going on of an ‘activity’ are the same type of going
on, i.e. roughly speaking, whether accomplishments are just ‘activities with boundaries.’
25 The neutral aspect consists in the lack of aspectual information, cf. Smith 1997. The
perfect (not to be confused with the perfective aspect!) is a mixture of tense and aspect; the
experiential perfect indicates that an occurrence has happened in the past and is over, the
resultative perfect conveys that an occurrence has happened in the past but has still current
relevance (compare the difference between Bill has been to America and Bill has gone to
America, cf. Comrie 1976).
26 Note that this feature provides some difficulty for those interpretations of Aristotle’s
energeia that take the latter to be ‘physis’, that which constitutes the living unity of an
organism, and present physis as a special kind of activity. On this issue cf. Gill in this
volume.
27 Cf. Kathleen Gill 1993.
28 Less than four-dimensional homeomerous entities are, for example, sensory impres-
sions and unbounded or bounded surfaces or boundaries. Among concrete entities, there
is no ‘contrast class’ to homeomerous entities; the purpose of such a general definition of
homeomerity is obviously to allow for a unified perspective of comparison.
29 Cf. Rijkhoff 2002. To highlight the connection to masses, I used to refer to free processes
as ‘dynamic masses.’
30 In previous expositions of FPT I failed to distinguish clearly between homeomerity and
automerity, and characterized free processes nominally merely as ‘homeomerous’ entities.
31 Cf. Seibt 2001b.
32 Cf. Heller 1990, Needham 1999, Rescher 2000, Stout 1997, Whitehead 1928, Zemach
1970. That processes are non-particular has been (implicitly) suggested first in Sellars
1981 (or so I argue in Seibt 2000b) and I myself have promoted this idea explicitly in
Seibt 1990b, 1995, 1997a, 1999, 2000d, 2001a,b. In his contribution to this volume Paul
Needham stipulates that processes have unique temporal occurrence in time but are not
located (i.e., also not multiply located) in space. This is a third position between the clas-
sical particularist approach to process ontology and the non-particularist approach of Free
Process Theory. Also, note that in the approach of Heller/Herre in this volume projections
of processes (onto portions of space and time) are particulars; I am unclear on whether
processes themselves are particulars in this framework.
33 Seibt forthcoming, ch. 3.
34 The following two sections contain an (I hope) improved version of an earlier exposition
of FPT (then called ‘APT’) in Seibt 2001.
35 Cf. Simons 1987.
36 For a first sketch of the formal framework FPT cf. Seibt 1990b; 1995, 1996c, 1999,
2000d, and 2001 contain versions of FPT at different stages of development.
37 For example, in [1] and [2] sentences (a) through (b) or (a) through (b), respectively,
do not entail sentence (C). [1] (a) Changing diapers is part of being a parent. (b) Opening
the box with wipes is part of changing diapers. (c) Pressing your thumb upward is part of
opening the box with wipes. (C) Pressing your thumb upward is part of being a parent. [2]
(a) The door is part of a house. (b) The hinge is part of the door. (c) The hinge is part of a
house.
38 Moreover, it also offers economical ways to model our reasoning about part-whole
relationships in artifacts pertaining to function, construction, material constitution, and
maintenance. For the latter compare Simons/Dements (1996).
39 To appreciate the usefulness of such a schematic approach to identity consider the fol-
lowing example, involving partitions on actions: If Kim plays violin at 9am at home –
she performs her entry exam to the conservatory at home at 9am – she is awakening her
neighbor Tom (a nocturnally working philosopher). Assume that Kim intended to perform
the entry exam and by doing so, to disturb her neighbor Tom. Performing an entry exam
is to play the violin in a place where an examiner is listening. If we define the identity
of actions in terms of 1-parts, direct intended consequences then the making music today
is identical to the performance of the entry exam but not to the disturbance of Tom. In
legal contexts, however, we might choose to define the identity of actions in a more fine-
grained fashion, e.g., in terms of an action’s 1-part and 2-parts, its instrumentally intended
consequences. In this case Kim’s playing on this occasion will be identical with disturbing
Tom on this occasion. In other words, FPT’s identity condition is as flexible as our actual
reasoning requires.
40 For example, the following two theorems do not hold (with ‘∩’ and ‘∪’ as abbreviations
for ‘prod’ and ‘sum’): (1) (w (x ∩ y) & v (x ∩ y) & w ◦ v) → (w ∩ v) (x ∩ y). (2)
(w (x ∪ y) & v (x ∪ y)) → (w ∪ v) (x ∪ y).
41 It holds that w (x ∩ y) ↔ w x & w y. Thus x ∩ (y ∩ v) = ιz (∀w (w x & w
(y ∩ v) → w z) = ιz (∀w (w x & w y & w v → w z) = (x ∩ y) ∩v. Similarly
for the sum operation because of x ◦ (y ∪ z) = x ◦ y ∨ x ◦ z.
42 I am assuming here that spacetime (either of physics or of common sense) is nonhomo-
geneous in the sense that different regions of spacetime are qualitatively different and
thus, according to the (P P P ) (i.e., Ax3), different spacetimings. A notational convention:
52 JOHANNA SEIBT
Variables in lower case Greek range over items in the field of ‘ ’ but (unlike ‘x, y, z’ etc.)
they are never co-denoting; variables ‘ρ’ with subscripts are reserved for ‘spacetimings.’
43 The last clause of (D9b) postulates that the instantiations of variable ‘x’ are not in
the ‘genus of spacetime’, i.e., that they are free processes which are not themselves
spacetimings. In addition, variables z in D9a and x in D9b, respectively, are restricted
to minimally self-contained processes. Note that minimal self-containment can be defined
without making explicit reference to part of spatiotemporal regions (as this was done above
for expository purposes): a minimally self-contained entity has no parts identical to itself.
44 An ultimate specific space-time region is a space-timing which is not a (functional, not
spatiotemporal!) part of any other space-timing.
45 Note that this is not the notion of ‘quantity’ used in Needham’s contribution to this
volume – quantities in Needham’s (Cartwright’s) sense are uniquely and determinately
located, i.e. they are particulars.
46 In the concluding paragraph of his contribution to this volume Wim Christaens abandons
(Ax5) and suggests that FPT could be used to talk about entities that do not exist in space
(as opposed to: do not have a determinate spatial location).
47 In the compressed presentation I can offer here it is easy to overlook the larger signific-
ance of the particularity theorem. In effect, FPT makes room for indeterminately located
indeterminate individuals and yet accommodates, as their limit case, determinately located
determinate individuals (particulars). This makes FPT a prima facie promising candidate
for the interpretation of quantum physical entities, cf. Seibt 2002 and Christiaens in this
volume.
48 Cf. Seibt 1996c.
49 With a unitary element characterized by dynamicity only (‘empty process’) and P being
the field of ‘ ’, the Structure W = P , sum forms an Abelian group. The definition of
‘dynamicity’ is presently under construction – in an earlier version of FPT (2001a) it is
defined as ‘dyn(x) ↔ ∃z, y (stage (z, x) & stage (y, x) & z¬y’ but this links dynam-
icity to change. In Seibt (forthcoming) I define dynamicity as self-production, inspired by
Aristotle’s characterization of energeia.
50 That spatiotemporal characteristics may be used to distinguish between basic ontolo-
gical categories (objects, events, properties, processes) is not new (cf. Mayo 1961 and,
closest in spirit to FPT, Zemach 1970). New is the idea that the latter can be considered as
species of one basic category.
51 For details cf. Seibt 2000d.
52 On transitivity alternations and ‘oblique subject’ alternations cf. Levin 1993: 25–45 and
79–82.
53 Events and properties may not be agents of one-agent processes, since sentences with
events and properties in subject position make implicit reference to other agents or patients.
This is partly obscured by lexical ambiguities, cf. (1) ‘the avalanche is racing down’ and
(2) ‘the avalanche came suddenly.’ In sentence (1) the subject term ‘avalanche’ denotes a
discrete expanse of matter, while in sentence (2) the subject term refers to an event, relating
its occurrence to an implicit patient for whom the occurrence was sudden.
54 Indexed quantifiers ‘∃1 ’, ‘∃2 ’, . . . etc. abbreviate the common definitions of numerals
by means of existential quantification. Subscripts of variables, e.g., ‘α1,2 ’, indicate type
restrictions on the quantification, relative to process types 1 through 5 as defined in 4.1.
55 For the latter compare for instance: ‘Lifting the lid only intensified the smell’, ‘the
tornado changed the color of the sky’.
56 For details compare Seibt 2001a, forthcoming.
57 Cf. Wilson/Lumsden 1991.
58 Cf. Christensen/Bickhard 2002.
59 Ibid.
60 In formulating these distinctions in dynamic context I take my bearings here from
Bickhard/Campbell 2000: 343.
61 (a) On individuation cf. Seibt 1995 and 1996a. (b) Since free processes can recur, the
qualitative identity of two regions can be accounted for in terms of the numerical identity of
a free process recurrent in both. Roughly speaking, ‘a is F and b is F’ is made true by three
processes, α, β, and γ , where γ is a +n-part of α and β. Cf. Seibt 1990a, 1990b, 2000b.
(c) The temporal recurrence of free processes opens up a new avenue to a solution to the
problem of persistence between endurance and perdurance, the so-called ‘recurrence view’
of persistence (cf. Seibt 1996c, 2003a). The recurrence theory of persistence endorses
the main tenet of the endurance theory that persistence is sameness over time. But such
sameness is not the numerical sameness of a particular, nor, as recently championed in
Simons 2000a, the sameness of a universal. This (Whiteheadian!) idea is already discussed
in Carter/Hestevold 1994 who – rightly I think – dismiss it as implying that our assertions
about persistence are about abstracta. FPT offers the option to think of such sameness as
the numerical identity of a concrete, recurrent individual.
62 Cf. Seibt 2000d and 2002; cf. also Christiaens’ contribution to this volume.
REFERENCES
Andersen, P. et al (eds.): 2000, Downward Causation: Mind, Bodies, and Matter, Aarhus:
University of Aarhus Press.
Bache, C.: 1995, The Study of Aspect, Tense and Action – Towards a Theory of the
Semantics of Grammatical Categories, Frankfurt a. M.: Peter Lang.
Bickhard, M. and D. Campbell: 2000, ‘Emergence’, in Anderson et al. 2000, 322–349.
Breu, W.: 2000, Probleme der Interaktion von Lexik und Aspekt (ILA), Tübingen:
Niemeyer.
Campbell, K.: 1990, Abstract Particulars, Oxford: Blackwell.
Christensen, W and D., M. Bickhard: 2002, ‘The Process Dynamics of Normative
Function’, The Monist 85, 3–28.
Carter, W. and H. Hestevold: 1994, ‘On Passage and Persistence’, American Philosophical
Quarterly 31, 269–283.
Comrie, B.: 1976, Aspect, Cambridge, UK: Cambridge University Press.
Dik, S.: 1997, The Theory of Functional Grammar, 2nd revised edition by Kees Hengeveld,
Berlin: Mouton de Gruyter.
Dowty, D.: 1979, Word Meaning and Montague Grammar, Dordrecht: Reidel.
Gill, K.: 1993, ‘On the Metaphysical Distinction Between Processes and Events’, Cana-
dian Journal of Philosophy 23, 365–384.
Gill, M.-L.: 1989, Aristotle on Substance. The Paradox of Unity, Princeton: Princeton
University Press.
Gill, M.-L.: 1994, ‘Individuals and Individuation’, in T. Scaltsas et al. (eds.), Unity, Identity,
and Explanation in Aristotle’s Metaphysics, Oxford University Press.
Heller, M.: 1990, The Ontology of Physical Objects, Cambridge, UK: Cambridge Univer-
sity Press.
Kearns, K.: 2000, Semantics, Houndmills: Macmillan.
54 JOHANNA SEIBT
Levin, B.: 1993, English Verb Classes and Alternations, Chicago: University of Chicago
Press.
Mayo, B.: 1970, ‘Objects, Events, and Complementarity’, Philosophical Review 70, 340–
361.
Mourelatos, A.: 1978, ‘Events, Processes and States’, Linguistics and Philosophy 2, 415–
434.
Needham, P.: 1999, ‘Macroscopic Processes’, Philosophy of Science 66, 310–331.
Pelletier, F.J.: 1974, ‘On Some Proposals for the Semantics of Mass Terms’, Journal for
Philosophical Logic 3, 87–108.
Pelletier, F.J.: 1978, Mass Terms. Some Philosophical Problems, Dordrecht: Reidel.
Rescher, N.: 2000, Process Philosophy, Pittsburgh: University of Pittsburgh Press.
Rijkhoff, J.: 1991, ‘Nominal Aspect’, Journal of Semantics 8, 291–309.
Rijkhoff, J.: 2002, The Noun Phrase, Oxford: Oxford University Press.
Rijksbaron, A.: 1989, Aristotle, Verb Meaning and Functional Grammar. Towards a New
Typology of States of Affairs, Amsterdam: J.C. Gieben.
Sasse, H.: 2002, ‘Recent Activity in the Theory of Aspect: Accomplishments, Achieve-
ments, or just Non-Progressive State?’, Linguistic Typology 6, 199–273.
Seibt, J.: 1990a, Properties as Processes. A Synoptic Study in W. Sellars’ Nominalism,
Reseda, CA: Ridgeview.
Seibt, J.: 1990b, Towards Process Ontology: A Critical Study in Substance-Ontological
Premises, Pittsburgh: University of Pittsburgh doctoral dissertation. Published as mi-
crofiche with UMI-Dissertation Publication, Michigan.
Seibt, J.: 1995, ‘Individuen als Prozesse: Zur prozess-ontologischen Revision des Substan-
zparadigmas’, Logos 5, 303–343.
Seibt, J.: 1996a: ‘Non-countable Individuals: Why One and the Same Is Not One and the
Same’, Southwest Philosophy Review 12, 225–237.
Seibt, J.: 1996b, ‘The Myth of Substance and the Fallacy of Misplaced Concreteness’, Acta
Analytica 15, 61–76.
Seibt, J.: 1996c, ‘Existence in Time: From Substance to Process’, in J. Faye, U. Scheffler,
and M. Urs (eds.), Perspectives on Time, Boston Studies in Philosophy of Science,
Dordrecht: Kluwer, pp. 143–182.
Seibt, J.: 1996d, ‘Der Umbau des “Aufbau”: Carnap und die analytische Ontologie’,
Deutsche Zeitschrift für Philosophie 44, 807–835.
Seibt, J.: 1997, ‘The “Umbau”: From Constitution Theory to Constructionism’, History of
Philosophy Quarterly 14, 305–351.
Seibt, J.: 1999, ‘Dinge als Prozesse’, in R. Hüntelmann and E. Tegtmeier (eds.), Neue
Ontologie und Neue Metaphysik, Köln: Academia Verlag, pp. 11–41.
Seibt, J.: 2000a, ‘Constitution Theory and Metaphysical Neutrality: A Lesson for Onto-
logy?’, The Monist 83, 161–183.
Seibt, J.: 2000b, ‘Pure Processes and Projective Metaphysics’, Philosophical Studies101,
253–289.
Seibt, J.: 2000c, ‘Ontology as Theory of Categorial Inference’, in D. Greimann
and C. Peres (eds.), Wahrheit–Sein–Struktur. Auseinandersetzungen mit Metaphysik,
Hildesheim: Olms Verlag, pp. 272–297.
Seibt, J.: 2000d, ‘The Dynamic Constitution of Things’, in J. Faye, U. Scheffler, and M.
Urchs (eds.), Facts and Events, Poznañ Studies in Philosophy of Science 72, 241–278.
Seibt, J.: 2001a, ‘Formal Process Ontology’, in C. Welty and B. Smith (eds.), Formal
Ontology in Information Systems: Collected Papers from the Second International
Conference, ACM Press: Ogunquit, 333–345.
Seibt, J.: 2001b, ‘Processes in the Manifest and Scientific Image’, in U. Meixner (ed.),
Metaphysik im postmetaphysischen Zeitalter/Metaphysics in the Post-Metaphysical Age,
Wien: öbv&hpt, pp. 218–230.
Seibt, J.: 2002, ‘Quanta, Tropes, or Processes: On Ontologies for QFT beyond the Myth
of Substance’, in M. Kuhlmann, H. Lyre and A.Wayne (eds.), Ontological Aspects of
Quantum Field Theory, Singapore: World Scientific, pp. 53–93.
Seibt J.: 2003, ‘Process and Particulars’, in M. Weber (ed.), Process Metaphysics, Munich:
Philosophia, forthcoming.
Seibt, J.: forthcoming, Free Process Theory: A Study in Revisionary Ontology, Habilita-
tionsschrift at the University of Konstanz, Germany.
Simons, P.: 1987, Parts: A Study in Ontology, Oxford: Oxford University Press.
Simons, P.: 2000, ‘Continuants and Occurrents’, The Aristotelian Society, Supplementary
Volume LXXIV, 78–101.
Simons, P. and C. Dements: 1996, ‘Aspects of the Mereology of Artifacts’, in R. Poli and
P. Simons, (eds.), Formal Ontology, Dordrecht: Kluwer, pp. 255–276.
Smith. C.: 1997, The Parameters of Aspect, Dordrecht: Kluwer Academic Publishers.
Stout, R.: 1997, ‘Processes’, Philosophy 72, 19–27.
Taylor, B.: 1985, Modes of Occurrence: Verbs, Adverbs, and Events, Oxford: Blackwell.
Whitehead, A.: 1928, Process and Reality, New York: Macmillan.
Wilson, E. and Ch. Lumsden: 1991, Biology and Philosophy 6, 401–412.
Zemach, E.: 1970, ‘Four Ontologies’, Journal of Philosophy 23, 231–247.

Typology of Occurings

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Typology of Occurings

Încărcat de

Drepturi de autor:

Formate disponibile

JOHANNA SEIBT

FREE PROCESS THEORY: TOWARDS A TYPOLOGY OF

The study of processes and process-based theory formation is a sur-

Axiomathes 14: 23–55, 2004.

posed, including Whitehead’s. Even though Free Process Theory operates

The task of ontology is to explore various ways of structuring the referen-

terms of L (e.g., ‘thing’, ‘property’, ‘person’ etc.). This explanation takes

1.1. ‘Action types’: an unsuccessful start

Kenny champions a threefold classification into ‘states’, ‘activities’ and

1.2. Aktionsarten and verbal aspect: aspectological lessons

‘Aktionsarten’ in terms of purely lexical criteria, i.e. in terms of possible

1.3. Activity as mode of occurrence

about aspectual implications can be put to good use. Instead of claiming

Since the clauses of these four definitions are conjuncted, an occurrence

(C7) Recurrence condition: According to (M3-iii), from ‘N V s’ we can

2. FREE PROCESSES : A NEW CATEGORY

If an occurrence occurs as activity, then it is conceived of as an entity

(HOM) Homeomerity: An n-dimensional (1 ≤ n ≤ 4) entity E of kind K

parts in dimension n are of kind K. In case (a) E is maximally likeparted,

An entity E that is homeomerous in this general sense can be the truth-

(AUT) Automerity: An n-dimensional (1 ≤ n ≤ 4) entity E of kind K

As witnessed by (C7) above, if we conceptualize an entity as an activity,

3. FPT: A MEREOLOGY ON FREE PROCESSES34

Classical Extensional Mereology (CEM) is geared to the reading of the

ticulating the fundamental sense of parthood should better be assigned to

Parthood-or-identity is then defined as

In FPT parthood as expressed by ‘ ’ reaches only into the immediate or

(D2) ∀n ∈ N; if n = 1 then x n y ↔ x y; if n ≥ 2 then x n y ↔

The expression ‘+m-part’ refers to all n-parts with m ≥ n. FPT is ‘ex-

(Ax3) ∀z(z n z↔z n y) ↔ x = y.

Since ‘ ’ is not the relation of extent part, the well-known counter-

(D3) x ◦ y ↔ ∃z (z x&z y),

but due to the non-transitivity of ‘ ’ the overlap of two entities we receive

(D5) x | y ↔ ¬∃z (z n x & z n y)

The supplementation principle in FPT is defined in terms of discreteness

This allows for wholes whose parts cannot be completely ‘disentangled’,

(D6) prod(x, y) =df ιz(∀w(w x&w y) ↔ w z)

(D7) sum(x, y) =df ιz(∀w(w ◦ z → w ◦ x ∨ w ◦ y))

On the basis of ‘≤’ the FPT-predicate ‘x is a phase in y’ can be introduced;

(D11) phase(x, y) ↔ x y & ∀z(z = [y] → ∃w (w = [x] & w < z)).

A free process α with phases can then be said to be a specification of a

(D12∗ ) specif(x, y) ↔ ∀z(phase(z, x) → z = [y])

If Max is a poodle, wherever there is an amount of being-Max, there is

(D13) ult-specif(x, y) ↔ specif(x, y) & ¬∃z(specif(z, x))

It is crucial to realize that in FPT an amount of a process α is a specification

(T1) ∀x [∃y ult-specif(x, y) → ∃z, ρi (x = I (z, ρi ) &

A third basic relationship between processes, besides phase and spe-

(D14) stage(x, y) ↔ ∀z(z = [x] → ∃w(w = [y] & z < w)

(Ax7) ∀x∃y(x y∨y x) → dyn(x)].49

4. TOWARDS A TYPOLOGY OF PROCESSES

4.1. Classification by automerity pattern

(1) Type 1 processes are temporally maximally self-contained and spa-

Kim’s pair-dancing). In the formal idiom of FPT the characteristic auto-

(2) Type 2 processes are temporally minimally self-contained while their

(3) Type 3 processes are spatially minimally self-contained and temporally

(4) Type 4 processes are spatially and temporally self-contained. Their

(5) Type 5 processes are spatially and temporally maximally self-

(20) α is type 5 iff ∀x (x <temp I (α, ρi ) → x = I (α, ρj )) & ∀x

It is important to note, however, that the conditions stated in (D16) through

4.2. Classification by dynamic parameter analysis

4.2.1. Participant structure

remained constant’, ‘this development could not be stopped’, etc. Some-

(1) A process α is a one-agent process iff there is an amount of α that is